The feral pigeon (Columba livia) is a globally commensal bird that can cause dis-amenities such as soiling and is a potential vector of various diseases. Aside to limiting food as a management strategy, reducing the availability of roosting and nesting sites can help regulate feral pigeon populations. Despite their prevalence, current knowledge of their roosting and nesting preferences is lacking.

Feral pigeons commonly use railway viaduct expansion gaps in Singapore for roosting and nesting. These gaps provided an ideal experimental platform to examine feral pigeon roost and nest site selection while controlling for differing cavity sizes which can significantly affect their reproductive success and site selection decisions. We also conducted an in-situ experiment to test the efficacy of nest removal as a management option.

Our nationwide surveys of 80.3km of railway viaducts and 6,048 gaps revealed that feral pigeon day roosting and nesting preferences are influenced by structural height and more importantly, their proximity to human food sources. There was a significantly higher probability of feral pigeon roosting in a gap if it had more pigeon feeding incidences in its vicinity and was higher. The probability of feral pigeon nesting in gap was higher if it was closer to a railway station, lower and further from water bodies.

In our field experiment, we did not find any significant differences in the proportions of the abundances of feral pigeon to other urban commensal bird species at the gaps before and after nests were removed.

Overall, our results suggest that a concerted effort to reduce anthropogenic food availability to feral pigeons is central in limiting their reproductive success and controlling their population.

We conducted our surveys on bird roosting and nesting along all above-ground MRT tracks on the main island of Singapore (1°21’N, 103°49’E) from January to February 2020 and repeated our surveys in March 2020. Nest surveys were not conducted in the subsequent survey as the occurrence of new nests between February and March 2020 is expected to be minimal. Our surveys were conducted during the dry phases of the Northeast Monsoon with less rainfall and coinciding with the start of the breeding season for resident birds in Singapore. In total, 80.3km of North-South and East-West viaducts amounting to 6,048 gaps, were surveyed. For each viaduct gap, we measured the height of the gap from the ground (mean, 6.33m), and recorded the state of deterrent measure used at the gap (‘absent’, ‘present’, or ‘damaged’), as well as the presence or absence of nesting materials and soiling as an indication of bird usage. We also noted the bird species observed in the gap. Surveys were conducted in clear weather (i.e., non-rainy days) from 0900 to 1900h. For gaps that were inaccessible because the viaduct cut across construction sites or private properties, we recorded the height of the gap only, and bird species occurring at the gap (~4% of the total number of gaps).

We observed a total of 2,264 columns and 6,048 viaduct gaps in 80.3 km of above-ground viaducts. For consistency, we only used 1,751 columns that supported four expansion gaps for analyses. To minimise spatial autocorrelation, we selected 230 columns that were spaced at least 100m apart from each other, beyond the feral pigeon natal distance of 90m (i.e., maximum distance travelled by juvenile birds from their nests). Only columns that did not have any deterrent installed were selected because they may exclude pigeons from using the gaps. We aggregated the total number of feral pigeons detected in the four gaps of each column to obtain abundance since these gaps were connected. Feral pigeon day roosts were defined as gaps where the presence of a feral pigeon was recorded in least one of our two sampling occasions (i.e., January-February 2020 and March 2020), and showed an indication of nesting or soiling to preclude gaps that were only used only momentarily (e.g., as a hideout from a potential threat). The presence of nesting materials in the viaduct gaps such as twigs and leaves were used as an indication feral pigeon nesting. Any nests that included the presence of bird species other than the feral pigeon were omitted from analysis on feral pigeon nesting preferences.

We selected the following environmental and structural factors that may influence day roost and nest site selection of pigeons along the viaduct:

(a) Intensity of pigeon feeding: “Mercy feeding” is a significant food source of pigeons in several cities, including Singapore. Data for public feedback cases on mercy feeding (i.e., incidences where intentional feeding of pigeons by people occurred and were reported by feedback providers) were obtained from the National Parks Board (NParks) from January 2019 to January 2020 (n = 3130) and aggregated within a 300 m radius of each column.

(b) Distance to nearest hawker centre: Hawker centres are food centres housed in open-air complexes made up of several small stalls selling cooked food. Urban commensal birds such as the house crow (Corvus splendens), Javan myna (Acridotheres javanicus) and feral pigeon have been found to be attracted to food wastes in the hawker centres. Here, we recorded the distance of each column to their nearest hawker centre.

(c) Distance to nearest water body: Water is essential to feral pigeons, particularly during breeding for the production of crop milk. Feral pigeons were also observed to nest and roost near water points in cities. Water bodies in our study include rivers, reservoirs, and large canals.

(d) Distance to nearest MRT station: MRT stations often have food outlets and their refuse collection points could also serve as a food source for feral pigeons. Thus, we calculated the distance from each column to the nearest MRT station.

(e) Mean distance to nearest five public residential buildings: We chose five buildings to better represent a collective contribution at the estate level as opposed to an individual building. Tall buildings are favoured by feral pigeons for roosting or nesting on the ledges or roofs to avoid predation and disturbances. Tall buildings were also highly correlated to pigeon density in Singapore. Here, we focused on public housing buildings, which are homes for most Singaporeans and typically range between 20 to 40 storeys high. We calculated the mean distance to the nearest five public housing buildings as our covariate.

(f) Height of MRT viaduct expansion gap: A measure of the structural height of the available nesting or roosting sites. Feral pigeons tended to nest or roost higher to avoid predator and other disturbance from the ground.

(g) Observation time: We calculated the minutes passed from our earliest observation, which was at 0900h, to the time each gap was observed.

The data files can be accessed by using the following softwares: Microsoft Excel or R Studio.